Ads in one category look the same because brands copy each other when nobody is sure what works. Economists call it herd behaviour and sociologists call it mimetic isomorphism: each brand follows its rivals and the category leader because that feels safer than an untested idea of its own. Each copy is defensible on its own, but together they remove the differences buyers use to choose. The way out is evidence rather than bravery: measure where your category converges, then test the gaps.
Key takeaways
- Herd behaviour can be rational for each brand and costly for all of them: economic models from 1992 show people following others even when their own information points elsewhere.
- In SOMIN's analysis of 637 labelled ads from 12 social-listening brands, 89% led with product features or an offer and 74% used a confident tone; testimonials appeared in about 1%.
- The same brands' organic posts were less uniform, with 44% leading on product, which suggests paid media adds its own pressure towards sameness.
- Not all imitation is a mistake: shared category cues help buyers recognise what you sell.
- To break away without guessing, measure convergence, pick gaps with real demand, change one thing at a time and pre-test before you spend.
What is herd behaviour in advertising?
Herd behaviour is doing what others do instead of acting on your own information. In advertising it shows up as brands adopting the messages, tones and formats their competitors already run.
The economics is well established. Abhijit Banerjee's 1992 paper "A Simple Model of Herd Behavior", in the Quarterly Journal of Economics, shows that when people can see earlier decisions, following them can be rational for each person, because the others may know something. The result, though, is an inefficient outcome for everyone. In the same year, Sushil Bikhchandani, David Hirshleifer and Ivo Welch described the mechanism as an informational cascade in the Journal of Political Economy: people stop using their private information and copy, and the resulting conformity is fragile, easily reversed by a little new public information.
Sociology describes the same thing at the level of organisations. In "The Iron Cage Revisited" (American Sociological Review, 1983), Paul DiMaggio and Walter Powell argued that organisations in one field grow more alike through three pressures: coercive (rules and powerful partners), normative (shared professional training and people moving between firms) and mimetic. Mimetic isomorphism is copying under uncertainty. In their words, "Uncertainty is also a powerful force that encourages imitation." Stanford GSB's 2024 retrospective on the paper summarises all three pressures.
Strategy research adds a useful split. Marvin Lieberman and Shigeru Asaba, in "Why Do Firms Imitate Each Other?" (Academy of Management Review, 2006), sort imitation into two families: information-based, where firms follow others they believe know more, and rivalry-based, where firms copy to keep up with rivals or to limit competition. Advertising has both. Brands copy the leader's message because it looks proven, and copy each other's offers so as not to fall behind.
How similar are ads in one category, really?
Very similar. We measured it in a category we track closely: social listening and social media management software. SOMIN's workspace follows 19 brands, SOMIN and 18 vendors. Between 1 July and 17 October 2026, 13 of them ran 840 ads that we collected from Meta's Ad Library (266) and Google's ads transparency data (574). SOMIN's models could label the tone, customer need and format of 637 of those ads, from 12 advertisers. We compared them with the same brands' organic posts and with ads from a second category, ad optimisation platforms.
| What the content leads with | Social-listening ads (637 ads, 12 advertisers) | Same brands' organic posts (1,730 posts, 16 brands) | Ad-optimisation ads (55 ads, 3 advertisers) |
|---|---|---|---|
| Product features or an offer | 89.2% | 44.2% | 98.2% |
| Confident tone | 73.5% | 64.5% | 90.9% |
| Informative tone | 77.7% | 70.9% | 52.7% |
| One of the set's two most common customer needs | 92.2% (knowledge, empowerment) | 90.1% (knowledge, growth) | 92.7% (growth, efficiency) |
| Testimonials and reviews | 1.1% | 5.4% | 0% |
Source: SOMIN analysis of ads and organic posts with readable labels, 1 July to 17 October 2026. Labels are inferred by SOMIN's models and one item can carry several. Ad-optimisation ads come from Smartly (33), Madgicx (20) and Strike Social (2).
Four things stand out:
- Ads are product-led almost without exception. All eight social-listening advertisers with at least ten labelled ads put product features or an offer in at least 65% of them, and most did so in more than 80%.
- Tone barely varies. Confident and informative tones dominate, and 52.4% of labelled ads carried both at once. Humour and playfulness appeared in one ad each out of 637.
- Customer proof almost disappears. Testimonials and reviews were in 7 of 636 ads with a labelled format, against 5.4% of organic posts.
- Needs converge in both channels. Nine in ten ads and posts speak to the set's two most common needs. The difference between paid and organic is in how they speak (product claims, confident tone), not what they speak about.
The second category tells the same story, with a caveat. 54 of 55 labelled ad-optimisation ads led with product, but that sample is small and comes mostly from two advertisers, so read it as a check rather than a finding.
Why do ads converge more than organic posts?
Our data shows the difference, not its cause. The research above points to pressures that are stronger in paid media than in organic posting:
- Uncertainty with money attached. Every ad has a cost and a short-term target, which rewards the choice that looks proven. That is DiMaggio and Powell's mimetic pressure in its purest form.
- Visible competitors. Public ad libraries make every rival's creative easy to study. That helps research, and it also makes copying effortless, the information-based imitation Lieberman and Asaba describe.
- Shared playbooks. Advertisers read the same platform guidance, agency templates and benchmark reports, and staff move between the same firms. That is the normative pressure.
- Parity anxiety. When a rival leads with a feature or a discount, matching it feels necessary, which is rivalry-based imitation.
- The same yardstick. When every advertiser tunes creative towards the same short-term metric, they tend to arrive at similar answers.
What does SOMIN's herd-behaviour insight say?
SOMIN generated an insight on this exact problem from its audience-research workspace, which tracks Nielsen, GWI, Milieu and Euromonitor. A study of 191 organic posts those firms published over two years produced a pain-point theme of 65 posts, 57 of them from Nielsen, about brands wasting spend through poor audience understanding and outdated assumptions. The insight built on it uses the angle "Herd Behavior in Advertising Trends", grounded in herd behaviour theory: brands follow prevailing trends without checking whether they fit their own audience, and the campaigns fail to connect. Its line is:
To truly connect, dare to diverge from the trending path.
An emotional moment generated in SOMIN's ad-optimisation workspace says the same thing from the inside: "I should have tailored our approach instead of applying the same strategy everywhere." Both are model output built from published content, not quotes from people. They show that the tension is visible in what the industry itself publishes.
Is following the herd always a mistake?
No. Some sameness is useful. Buyers rely on category cues to recognise what a product is, and an ad that breaks every convention may simply not be understood. Matching a rival's offer can be rational, as rivalry-based imitation predicts. And a herd can be right: if every competitor leads with a feature because buyers really do choose on it, ignoring that feature is not bold, it is wrong.
The cost appears at category level. When nearly every ad makes a confident product claim, no single claim stands out, and buyers fall back on familiarity and price. The useful distinction is between codes that tell buyers what the category is, which you keep, and choices that tell them who you are, which you should not copy.
AI makes the problem faster but did not create it. We cover the AI-specific version, similar models fed similar briefs, in AI content homogenisation. Herd behaviour predates generative tools and would exist without them.
How do you break from the herd without guessing?
Treat difference as a hypothesis to test, not a leap of faith.
- Measure convergence. Collect a quarter of your category's ads and organic posts, label each for need, tone, format and audience, and calculate each label's share and the number of brands using it. Our guide to researching competitor ads in the Meta Ad Library covers the collection step.
- Shortlist gaps with demand. In the social-listening data, testimonials (1.1% of ads), trust as a customer need (3.6% of ads) and humour (two ads) are barely used. Low use is not proof of opportunity, so check audience demand first, as described in what marketing whitespace is.
- Keep the category cues and change one thing. Keep whatever tells buyers what you sell. Change the variable you are testing, such as the type of proof, the tone or the audience, so you learn what made the difference.
- Pre-test before you spend. Score the herd version and the different version for predicted performance before launch. SOMIN's ad scoring predicts click-through from what wins in your category.
- Run a controlled test. Give both versions the same budget, audience and placement, with one primary metric agreed in advance.
- Re-measure every quarter. Herds move. If a gap you open starts attracting followers, that is the clearest sign it worked, and the cue to look for the next one.
FAQ
What is herd behaviour in marketing?
It is following what competitors or the market do instead of acting on your own evidence. Economists have shown it can be rational for each decision-maker under uncertainty, while producing a worse outcome for the group as a whole.
What is mimetic isomorphism?
It is the sociologists' term, from DiMaggio and Powell's 1983 paper, for organisations copying similar organisations when they are unsure what works. It is one of three pressures that make organisations in a field more alike, alongside coercive and normative pressure.
Why do B2B software ads look so similar?
In our analysis of social-listening software, most ads led with product features in a confident, informative tone and spoke to the same two customer needs. Uncertainty about what works, easy access to rivals' ads and shared playbooks all push in that direction.
How can I tell whether my category's ads are converging?
Label a quarter of competitor ads for need, tone, format and audience, then check how much of the total the top two or three labels cover and how many brands use them. If a few labels cover most ads across most brands, the category has converged.
Is it risky to look different from competitors?
It can be, which is why the difference should be tested rather than assumed. The risk of sameness is quieter but real: spend that buys attention for the category instead of for your brand.
Sources
- Abhijit V. Banerjee, "A Simple Model of Herd Behavior", Quarterly Journal of Economics 107(3), 1992, pp. 797–817.
- Sushil Bikhchandani, David Hirshleifer and Ivo Welch, "A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades", Journal of Political Economy 100(5), 1992, pp. 992–1026.
- Paul J. DiMaggio and Walter W. Powell, "The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields", American Sociological Review 48, 1983, pp. 147–160; summarised by Stanford GSB, April 2024.
- Marvin B. Lieberman and Shigeru Asaba, "Why Do Firms Imitate Each Other?", Academy of Management Review 31(2), 2006, pp. 366–385, listed on Lieberman's publications page.
